Archaeological Predictive Modelling Research Articles

The habitation Model Trend Calculation (MTC): A new effective tool for predictive modeling in archaeology

ABSTRACT The aim of this paper is to create and present a new archaeological predictive model via GIS, incorporating what archaeologists consider the most important criterion absent of similar past models, that of critical thinking. The new model suggested in this paper is named habitation Model Trend Calculation (MTC) and is not only integrates the archaeological questions with a critical view, but it can be easily adjusted, according to the conditions or the questions concerning the archaeological community. Furthermore, it uses new topographical and geomorphological indexes such as Topographical Index (TPI), Hillslope and Landform Classification that give a new sense of the topographical and geomorphological characteristics of the examined area; therefore this model is a more powerful tool compared to older models that did not use new topographical and geomorphological indexes. The success of the created model is checked as a case study in the region of Messenia, Greece during the Mycenaean era. The region of Messenia is considered as one of the most important Mycenaean regions of Greece due to the great number and the importance of Mycenaean sites identified. For the present paper, 140 habitation sites were divided into four hierarchical categories (centers, large villages, villages, and farms) based on the extent and the plurality of the tholos tombs that exist in the broader region and according to the hierarchical categorization used by the archaeologists who have studied the area. The new predictive model presented in this work can assist in solving a series of criticisms that have been expressed in the previous studies regarding such models. Additionally, in the case of Mycenaean Messenia, the model shows excellent results in relation to the habitats of the time.

GIS-based Landform Classification of Eneolithic Archaeological Sites in the Plateau-plain Transition Zone (NE Romania): Habitation Practices vs. Flood Hazard Perception

The landforms of the Earth’s surface ranging from large-scale features to local topography are factors that influence human behavior in terms of habitation practices. The ability to extract geomorphological settings using geoinformatic techniques is an important aspect of any environmental analysis and archaeological landscape approach. Morphological data derived from DEMs with high accuracies (e.g., LiDAR data), can provide valuable information related to landscape modelling and landform classification processes. This study applies the first landform classification and flood hazard vulnerability of 730 Eneolithic (ca. 5000–3500 BCE) settlement locations within the plateau-plain transition zone of NE Romania. The classification was done using the SD (standard deviation) of TPI (Topographic Position Index) for the mean elevation (DEV) around each archaeological site, and HEC-RAS flood hazard pattern generated for 0.1% (1000 year) discharge insurance. The results indicate that prehistoric communities preferred to place their settlements for defensive purposes on hilltops, or in the close proximity of a steep slope. Based on flood hazard pattern, 8.2% out of the total sites had been placed in highly vulnerable areas. The results indicate an eco-cultural niche connected with habitation practices and flood hazard perception during the Eneolithic period in the plateau-plain transition zone of NE Romania and contribute to archaeological predictive modelling.

Toward archaeological predictive modeling in the Bosutswe region of Botswana: Utilizing multispectral satellite imagery to conceptualize ancient landscapes

Geospatial technologies such as GIS, remote sensing, and GNSS are transforming landscape archaeology, particularly in research areas that are hard to access due to geographical size, environmental challenges, lack of infrastructure, and/or political instability. In the following article, we explore the potential of high-resolution, multispectral satellite imagery for building a predictive model of the settlement pattern of the Iron Age polity of Bosutswe (700–1650 CE) in Botswana. While the case study lies in Africa, the problems that Africanists experience extend far beyond the continent. With limited funding, poor access to survey regions, and countless others, predictive modelling has the potential to identify more robust archaeological landscapes.Using ArcGIS 10.6, ENVI 5.4, and WorldView 2 and 3 imagery, we catalog 22 prospective archaeological sites within a 5 km radius of Bosutswe (100 sq. km), demonstrating the potential for image classification to identify sites difficult to locate in a traditional pedestrian survey. The discovery of these new, often small sites has implications for past environmental utilization and for regional sociopolitical dynamics. We describe our test excavation at one new site, Letlalolanoga, to demonstrate how it may contribute towards a growing body of literature that favors heterarchical power relations within African complex societies.

Remote-sensing data-based Archaeological Predictive Model (APM) for archaeological site mapping in desert area, South Morocco

Morocco hosts numerous archaeological sites, some of which are part of the UNESCO world heritage. Many of these sites, especially funerary mounds also called tumuli, or rock engravings and ceramics, are located in remote areas with limited access, particularly in the Saharan Morocco desert. We developed a remote sensing and GIS model to identify areas with high potential for hosting archaeological sites in the Awserd region of southern Morocco. A field campaign in a “reference site” zone of 21km2 has revealed 233 archaeological sites. Here we use satellite images and Digital Elevation Models to examine with various techniques (spatial analysis, statistical techniques, and fuzzy logic functions) the relations between the distribution of the archaeological sites and geo-environmental variables such as ground geology, topographic elevation and slope, orientation (aspect), and distance to water sources. We derive empirical relations that reveal that the distribution of archaeological sites depends on the above geo-environmental variables. We then use the empirical relations to anticipate the potential locations of archaeological sites in a region of 980km2 enclosing the reference site area. The model proves capable of predicting 582 sites in the larger region. Subsequent field observations there confirmed that about 80% of the model anticipations were correct. Our Archaeological Predictive Model (APM) can be scaled to larger areas and varied geographic settings, and hence can be a useful guide for archeological studies in desert regions.

A Spatial Pattern Analysis of Frontier Passes in China’s Northern Silk Road Region Using a Scale Optimization BLR Archaeological Predictive Model

In China’s Northern Silk Road (CNSR) region, dozens of frontier passes built and fortified at critical intersections were exploited starting at approximately 114 B.C. to guarantee caravan safety. Understanding the pattern of these pass sites is helpful in understanding the defense and trading system along the Silk Road. In this study, a scale optimization Binary Logistic Regression (BLR) archaeological predictive model was proposed to study the spatial pattern of CNSR frontier passes for understanding the critical placement of ancient defense and trading pass sites. Three hundred and fifty sample locations and 17 natural proxies were input into the model. Four strongly correlated factors were reserved as independent variables to construct the model, which was validated by 150 surveyed data and Kvamme’s Gain statistics. According to the variable selection and model optimization, the best spatial scale varies with the stability of the variables, such as 50 m and 1000 m, respectively, for the terrain and non-terrain variables. Clustering characteristics were identified with division overlapped with a 400 mm precipitation line using the site sensibility map. The high and medium probability areas were assembled along the Great Wall and the CNSR routes, especially in the western part, revealing that the model is also helpful to reconstruct the Silk Road routes.

Predictive modeling for archaeological site locations: Comparing logistic regression and maximal entropy in north Israel and north-east China

Archaeological predictive modeling is a tool that helps assess the likelihood of archaeological sites being present at different locations in the landscape. Such models are used for research purposes, as an analytical tool to better explain settlement patterns and past human behavior. They are also an important tool for the preservation of archaeological sites, as they can help planners avoid areas where sites are likely to exist. In this study we compare two methods of predictive modeling for archaeological site locations using two independent case studies. The more commonly used method of logistic regression is compared with a newer method of maximal entropy (MaxEnt). We examine the effectiveness of both models on two independent datasets collected from the Upper Galilee (northern Israel) and the Fuxin area (northeast China). While both methods have proven useful, in both cases the MaxEnt models produced much better results, which were much more efficient, than those of the logistic regression.

GIS-based landform classification of Bronze Age archaeological sites on Crete Island.

Various physical attributes of the Earth’s surface are factors that influence local topography and indirectly influence human behaviour in terms of habitation locations. The determination of geomorphological setting plays an important role in archaeological landscape research. Several landform types can be distinguished by characteristic geomorphic attributes that portray the landscape surrounding a settlement and influence its ability to sustain a population. Geomorphometric landform information, derived from digital elevation models (DEMs), such as the ASTER Global DEM, can provide useful insights into the processes shaping landscapes. This work examines the influence of landform classification on the settlement locations of Bronze Age (Minoan) Crete, focusing on the districts of Phaistos, Kavousi and Vrokastro. The landform classification was based on the topographic position index (TPI) and deviation from mean elevation (DEV) analysis to highlight slope steepness of various landform classes, characterizing the surrounding landscape environment of the settlements locations. The outcomes indicate no interrelationship between the settlement locations and topography during the Early Minoan period, but a significant interrelationship exists during the later Minoan periods with the presence of more organised societies. The landform classification can provide insights into factors favouring human habitation and can contribute to archaeological predictive modelling.

Using satellite image analysis for locating prehistoric archaeological sites in Alaska's Central Brooks Range

In this pilot study, we apply satellite image analysis to archaeological site prospection in Alaska's Brooks Range. Our goal was to test whether satellite remote sensing, which has been successful in locating large archaeological features associated with sedentary peoples, could be applied to arctic interior sites associated with mobile hunter–gatherers. In particular, we strove to develop a relatively straightforward and inexpensive model using existing data which could be used to help guide archaeology surveys. Using 1-m resolution IKONOS imagery of Lake Matcharak along the upper Noatak River, we produced a Normalized Difference Vegetation Index (NDVI) and tasseled cap transformation of the study area and stacked these five vegetation indices into one image. We then applied unsupervised and supervised classifications to the image first, to test for the presence of a site-specific spectral class and second, determine the nature of that class. Through a visual analysis of the unsupervised classification, a spectral phenomenon was seen to co-occur with archaeological sites in the study area. The supervised classification provided a high-resolution land cover map used to identify the signature as the ecotone between un-vegetated sediments and dense willow (Salix sp.) stands. Dense willow stands along the Lake Matcharak shore visually correlate with most of the known archaeological sites, possibly reflecting landform and/or vegetation characteristics that would have appealed to past inhabitants. The methods described here could contribute to building better survey strategies and archaeological predictive models for elsewhere in the Brooks Range and Alaska.

Archaeological Predictive Model – Manning Diversified Forest Products Ltd., P15 Forest Management Unit

This document describes an archaeological predictive model prepared for Manning Diversified Forest Products Ltd. (MDFP) of Alberta. The model refers to the P15 Forest Management Unit (FMU) and is for use in conjunction with their Historical Resource Management System. Multi-criteria evaluation was used to create the model. The modeling process involved the incorporation of data layers including Alberta Vegetation Inventory (AVI) layers from MDFP, a 25-m digital elevation model (DEM), topographic features, a cost raster of values indicating ease of travel, and terrain roughness. The FMU was stratified into two areas to ensure greater internal similarity. The model identified high-potential areas, in which archaeological sites can be expected, and no-potential areas, in which they would be expected to be absent. The next step should be to test the accuracy of the model. Following an initial testing period, changes may be made to increase precision. Increases in accuracy will result from verifications in the accuracy of the predictions, archaeological sites being found in areas identified in the model as having high potential and no sites found in areas of no potential. Increases in the accuracy of the model will also result with the incorporation of better resolution base mapping data.

Selected Mathematical Principles of Archaeological Predictive Models Creation and Validation in the GIS Environment

Selected Mathematical Principles of Archaeological Predictive Models Creation and Validation in the GIS Environment

Qualitative Landscape Theories and Archaeological Predictive Modelling—A Journey Through No Man’s Land?

An approach to test and interpret geographic information system-based, quantitative archaeological predictive models using techniques from qualitative (experiential) landscape theory is described and demonstrated via a case study. The result is the transformation of statistical output from quantitative predictive models into more archaeologically meaningful interpretations of spatial data.

MODELLING VAGUE KNOWLEDGE FOR DECISION SUPPORT IN PLANNING ARCHAEOLOGICAL PROSPECTIONS

Abstract. Most archaeological predictive models lack significance because fuzziness of data and uncertainty in knowledge about human behaviour and natural processes are hardly ever considered. One possibility to cope with such uncertainties is utilization of probability based approaches like Bayes Theorem or Dempster-Shafer-Theory. We analyzed an area of 50 km2 in Rhineland Palatinate (Germany) near a Celtic oppidum by use of Dempster-Shafer's theory of evidence for predicting spatial probability distribution of archaeological sites. This technique incorporates uncertainty by assigning various weights of evidence to defined variables, in that way estimating the probability for supporting a specific hypothesis (in our case the hypothesis presence or absence of a site). Selection of variables for our model relied both on assumptions about settlement patterns and on statistically tested relationships between known archaeological sites and environmental factors. The modelling process was conducted in a Geographic Information System (GIS) by generating raster-based likelihood surfaces. The corresponding likelihood surfaces were aggregated to a final weight of evidence surface, which resulted in a likelihood value for every single cell of being a site or a non-site. Finally the result was tested against a database of known archaeological sites for evaluating the gain of the model. For the purpose of enhancing the gain of our model and sharpening our criteria we used a two-step approach to improve the modelling of former settlement strategies in our study area. Applying the developed model finally yielded a 100 percent success rate of known archaeological sites located in predicted high potential areas.

Integrating Archaeological Theory and Predictive Modeling: a Live Report from the Scene

Archaeological predictive modeling has been used successfully for over 20 years as a decision-making tool in cultural resources management. Its appreciation in academic circles however has been mixed because of its perceived theoretical poverty. In this paper, we discuss the issue of integrating current archaeological theoretical approaches and predictive modeling. We suggest a methodology for doing so based on cognitive archaeology, middle range theory, and paleoeconomic modeling. We also discuss the problems associated with testing predictive models.

Predictive Modelling and Time: An Experiment in Temporal Archaeological Predictive Models

One of the most common criticisms of archaeological predictive modelling is that it fails to account for temporal or functional differences in sites. However, a practical solution to temporal or functional predictive modelling has proven to be elusive. This article discusses temporal predictive modelling, focusing on the difficulties of employing temporal variables, then introduces and tests a simple methodology for the implementation of temporal modelling. The temporal models thus created are then compared to a traditional predictive model.

Identifying at-risk heritage resources with GIS: modelling the impact of recreational activities on the archaeological record

Archaeological predictive models were initially developed to determine the probability of encountering certain types of archaeological sites (e.g. masonry ruins or artifact scatters) on a given unit of land. Other than forecasting densities of archaeological site-types by vegetation zones, however, such studies rarely focused on management applications and, consequently, were nearly abandoned. To illustrate the role that GIS can have in assessing and managing modern threats to heritage resources, we explore the usefulness of a predictive model that determines the extent to which off-highway access points affect the likelihood that heritage resources will be impacted by camping, hunting or woodcutting. With high-intensity GPS-based survey data from the Upper Basin, northern Arizona, this study shows how the capabilities of GIS may be extended to assist heritage managers and researchers in their efforts to identify areas where archaeological resources should be protected from various kinds of recreational activities on public lands.

‘Belief’ in the past: Dempster-Shafer theory, GIS and archaeological predictive modelling

This paper introduces the use of a new technique in archaeological predictive modelling which has particularly wide ranging appeal and application in the cultural heritage management sphere. Most predictive modelling programs are restricted to the use of certain types of data with a range of untested assumptions or caveats. Traditional probability based models are difficult to construct, and are often limited by the use of data sets which render the results statistically suspect. The method introduced here makes use of non-parametric statistical methods which are not hampered by imperfect raw data. In particular, this paper introduces the use of Dempster-Shafer theory in archaeological predictive modelling

Archaeological predictive modelling in raised mires – concerns and approaches for their interpretation and future management

Whilst raised mires retain high potential for exceptional preservation of organic archaeological remains, locating such features is problematic. Within dryland landscapes, Geographical Information Systems (GIS) provide a range of tools that enable the predictive modelling of site locations on the basis of observed trends in the relationships between environmental conditions and human agency. However, the formation processes of raised mires mean that this is apparently not feasible due to considerable spatial and temporal complexity. This paper re-defines this position arguing that predictive modelling of archaeological site locations within raised mires is possible through four-dimensional landscape reconstruction and analysis, using multiple strands of archaeo-environmental data.

Application of Landsat Data and Synoptic Remote Sensing to Predictive Models for Prehistoric Archaeological Sites: An Example from the Delaware Coastal Plain

In recent research using prehistoric archaeological site data from the Delaware Coastal Plain we have developed techniques to apply Landsat remote sensing data to the generation of archaeological predictive models. Rather than focusing on the specific sensing of archaeological sites, we use a synoptic approach that delineates the likely environmental settings for archaeological sites. The logistical regression statistical technique provides a quantitative link between site locations and environmental variables. By analyzing the environmental variables associated with known site locations, and known non-site locations, the logistical regression provides a probability assessment of an unsurveyed area's potential for containing archaeological sites. Supervised classification of Landsat data generates maps of environmental zones that can be related to the environmental variables used in the logistical regression. The model and the Landsat classifications have been subjected to preliminary tests and they provide accurate assessments of site potential in unsurveyed areas, as well as measures of the accuracy of the assessments. Results can be applied to cultural resource management problems and also yield useful data on prehistoric land use patterns.